lucid-fdn/backtesting

Backtesting

Test trading strategies against historical OHLCV data to evaluate performance before risking real capital.

SMA Crossover Strategy

A trend-following strategy based on moving average crossovers.

Rules

• Entry (Golden Cross): Fast SMA crosses above Slow SMA (previous bar: Fast <= Slow, current bar: Fast > Slow)
• Exit (Death Cross): Fast SMA crosses below Slow SMA (previous bar: Fast >= Slow, current bar: Fast < Slow)
• Direction: Long only (buy on golden cross, sell on death cross)

Default Parameters

| Parameter | Default | Description | |-----------|---------|-------------| | Fast Period | 10 | Fast SMA lookback | | Slow Period | 30 | Slow SMA lookback | | Minimum Bars | 500 | Minimum OHLCV bars needed |

How It Works

1. Compute SMA(fastPeriod) and SMA(slowPeriod) on close prices
2. Align arrays: offset = slowPeriod - fastPeriod
3. Walk through bars starting from slowPeriod:
   • If fast SMA crosses above slow SMA and no open position: BUY at close price
   • If fast SMA crosses below slow SMA and position is open: SELL at close price, record trade PnL
4. Calculate performance metrics on all completed trades

See references/strategies.md for full pseudocode.

RSI Mean-Reversion Strategy

A mean-reversion strategy that buys oversold conditions and sells overbought conditions.

Rules

• Entry: RSI crosses above the oversold level (previous bar: RSI <= oversold, current bar: RSI > oversold)
• Exit: RSI crosses above the overbought level (previous bar: RSI <= overbought, current bar: RSI > overbought)
• Direction: Long only

Default Parameters

| Parameter | Default | Description | |-----------|---------|-------------| | RSI Period | 14 | RSI calculation lookback | | Oversold Level | 30 | Entry trigger level | | Overbought Level | 70 | Exit trigger level |

How It Works

1. Compute RSI(period) on close prices
2. Walk through RSI values starting from index 1:
   • If RSI crosses above oversold threshold and no open position: BUY at close price
   • If RSI crosses above overbought threshold and position is open: SELL at close price, record trade PnL
3. Calculate performance metrics on all completed trades

See references/strategies.md for full pseudocode.

Backtest Engine

Equity Calculation

Starting equity: $10,000

equity = 10000
For each completed trade:
  equity = equity * (1 + trade_pnl_pct / 100)
Total Return % = ((final_equity - 10000) / 10000) * 100

This models fully invested equity (each trade uses 100% of current equity).

Result Metrics

Every backtest produces these metrics:

| Metric | Formula | |--------|---------| | Total Return % | ((Final Equity - 10000) / 10000) * 100 | | Sharpe Ratio | Annualized mean return / annualized std dev (365-day) | | Max Drawdown % | Maximum peak-to-trough decline in equity series | | Win Rate % | Winning trades / total trades * 100 | | Profit Factor | Gross profit / gross loss | | Total Trades | Count of completed round-trip trades |

Plus the full trade list with entry/exit prices, times, and PnL for each trade.

Additional Strategy Configs

DCA (Dollar-Cost Averaging)

| Parameter | Description | |-----------|-------------| | investmentPerTrade | Fixed USD amount per buy | | frequency | Interval between buys (e.g., "daily", "weekly", "monthly") |

DCA buys a fixed dollar amount at regular intervals regardless of price. Effective for reducing impact of volatility on average entry price.

Grid Trading

| Parameter | Description | |-----------|-------------| | bottomPrice | Lower bound of the grid | | topPrice | Upper bound of the grid | | gridLevels | Number of grid lines between bottom and top | | investmentAmount | Total capital allocated to the grid |

Grid spacing: (topPrice - bottomPrice) / gridLevels

Each grid level has a buy order (when price drops to that level) and a sell order (at the next level up). Investment per level = investmentAmount / gridLevels.